Streamline DDQs With AI: Faster Answers, Stronger Governance

Mark Nafash

00:04 - 00:23

Hello, everyone. Welcome to Streamline DDQs with AI, faster answers and stronger governance.

I'm Mark Mafesh with SS and C Intralinks. I'm a sales director on the alternative investment team.

And I lead our team here in North America focused on selling into our internal investment customers. Now I'm gonna pass it over to Maron to introduce herself.

Maren Fajardo

00:23 - 00:33

Thanks, Mark. Nice to meet you all.

My name is Lauren Fajardo. I'm a sales engineer here based out of New York, specifically focused in alternative investments.

Mark Nafash

00:33 - 04:29

Great. Thank you, Martin.

So today, we're here to talk about DDQs. I'm gonna go through a little bit of a background on SS and C Intralinks.

So SS and C Intralinks has an extensive track record of innovation since we started the company over thirty years ago. We started investing in AI in 2018 when we established our data science team in our AI practice internally at Intralinks with the goal of using AI to really enhance our products.

Since then, we've seen a lot of growth with our AI tools. We began integrate.

Integrating our AI features into our VDR in 2021, and we handpicked highly rated AI models and apply and transfer those techniques to customize in our product environment. In 2023, we launched Linq, our proprietary AI engine.

And last year, we integrated Linq into our Alts products starting with LPs, giving them ability to scrape data from their DP portals. We invested millions in AI over the last five years, and our AI models are purpose built for the process our customers perform in our products today, and we continue to invest in it every day.

FundCenter AI supports the entire fund life cycle. FundCenter is purpose built, an AI enabled solution that helps fund managers raise capital faster, onboard investors seamlessly, and simply.

And simplify investor reporting and communications all in a single platform. It allows you to scale your operations while providing an exceptional investor experience for your LPs.

FCAI leverages AI for the full investor journey. It gives your team a unified AI enhanced view of the full investor journey across your funds, starting with pre marketing, fundraising, onboarding, and reporting.

The analytics you see, you can see which investors are most engaged, where things are getting stuck, and where to focus your team's energy all without even leaving the platform. FundCenter AI's Link, our proprietary AI engine, is purpose built for alternative investments.

With Ask Link, your team can instantly locate critical information from thousands of documents just by asking questions. No manual searching.

Link pinpoints exactly what you need and generates concise summaries to accelerate and review both GPs and LPs within bank grade security. Now looking at DDQ, the goal of DDQ is to take you from weeks to minutes.

Our DDQ AI functionality leverages what's in your fundraise, leverages your existing documents to produce results and answer the questions that your LPs are asking. You know, knowledge base to draft and refined responses, we're really trying to enable you to, reduce the amount of time spending doing DDQs all within our platform.

LPs can upload their DDQ directly to the portal, and the platform tracks all versions automatically. And there's also a built in messaging to keep your team and LPs collaborating in real time.

The three pillars of our AI difference is really integrated fund management, AI powered with link, and simplified workflows. We're really focused on these three things to improve our AI and make things simpler for our clients, for the GPs and LPs moving forward within our product.

Now I'm gonna pass it over to Marin to do a demo on our GDQ platform.

Video

04:29 - 09:55

Thank you, Mark. As Mark mentioned, I will run through the demo of the DDQ.

So the Fund Center Fundraising Investor DDQ model is a high efficiency tool that really automates the most repetitive aspects of DDQ completion. So powered by our proprietary AI assistant, Ask Link, combines intelligent content retrieval to AI assisted drafting, workflow automation, and real time collaboration.

So it really is to enable you guys, our GPs, to complete those DDQs faster and more accurately. One thing we're gonna show here first is I'm coming in as you.

So as the GPS here, and I'm able to see this page of reference documents. This is found under documents, and this is just internal documentation that can be uploaded to the site with these AI insights turned on.

This just means that the AI, piece in here will only be reviewing these specific documents that have this turned on. So as I go over into workflow, this is actually where I can configure a workflow for my DDQ piece.

So as I can see here, there is an answer and a reviewer. This just means that there are specific roles within my internal team on the GP side that can actually answer those questions and then further review them for completion.

At the bottom, we can also see instructions for AI processing. Basically, this can just be firm specific terminology as I am answering those specific questions.

But we're gonna go over into the DDQ piece here, and this is where we can actually add DDQs from our investors. So they would send them to us.

These LPs would be sending them to us, and we're able to download what those templates may look like. I have one specifically uploaded already.

We're just going to upload that. I can rename this what I need to and then associate it with a specific investor.

From there, if there are multiple different vehicles that that investor is associated with, I can just choose one and then associate a due date with them. I already have this uploaded, so we're actually going to go over and review it in the a g two r investment fund.

From here, this is where I'm able to view whatever that 40 to 300 questionnaire may look like. From here, I can also see that it has been claimed already by me, but I can also move it for review once I am done with this document.

Before we get to that, I also want to show that there is internal commenting between team members so that there can actually be another piece of an audit trail so that you can communicate between your team members. But going over into the most important part here is the questions.

I'm actually able to see that each of these individual questions has already had a first pass taken by the AI tool. I can review where those reference documents have come from specific to the page number and if it has been manually overwritten.

So if I want to come in here and just reconfigure this sentence, or make it look just a little bit different, I can do that. I can also make the manual for the preferred answer.

Not only that, but I can regenerate these AI answers specific to individual documents. So if I know that this first question only needs to be answered by the business model and the DDQ piece, I could regenerate that answer.

Also, I can add more instructions for the AI processing for this specific question. This just gets into more granular detail as you're able to come in and review and repeat those questions.

From here, I can generate an answer file and choose a Word or Excel file and then save it in here. That will be showing up in the version control where I'm actually able to see each pass at the versions, basically, an audit trail to where these are.

Some of the key points in here that I want to mention is that it's only going to be answering questions based off of this designated document. So there's no learning within the AI tool, and it also will be accepting the PDF file starting last month.

From here, though, I would actually move it for review, and then the next team member could come in here and complete this file. So once that is reviewed and marked as completed, it cannot be touched again as a DDQ file.

I can also see the activity log in here. So what are my team members coming in and doing? Are they actually answering these specific questions and what questions have been answered? So we can see from here this has been manually overwritten from one thing to the next.

That will be the end of the demo, and I'm going to pass it back over to Mark, where we're happy to answer any questions. Thank you so much.

Mark Nafash

09:55 - 10:23

Thanks, Martin. So now we're gonna do some, questions.

I believe we already have a few. So I'm going to read the one in the chat, and I.

think Marin's gonna answer it. If there are documents already loaded in a VDR, can you pull from your FundCenter VDR so you do not upload the documents twice?

Maren Fajardo

10:23 - 10:38

It depends on where those documents are uploaded in your VDR. But with that AI insights piece that we saw, you can just turn that on, and that would prompt the AI tool to pull, from that specific document.

Mark Nafash

10:38 - 10:45

Great. Are there any other questions out there? Feel free to use the chat to ask any other questions, and we're happy to answer them now.

Maren Fajardo

10:45 - 11:14

It looks like there's one more above it. In the audit trail, is there a way to see track changes? Yes.

Correct. So that audit trail is gonna track every click that's taken in the data room.

You're actually able to go in and see what changes are being made in there. So, hopefully, that answers that question.

Mark Nafash

11:14 - 11:21

Great. Any other.

Oh, does this work with the LPADDQ template?

Maren Fajardo

11:21 - 11:28

Yes. It does work with the upload data to queue template.

Mark Nafash

11:28 - 11:35

Do users build a knowledge base or you always need to upload the docs? So I believe you already answered that one.

Maren Fajardo

11:35 - 12:21

Is that based off of. If that means, like, your documents will be learning or or the AI tool will be learning based off of of your documents, no.

So there's guardrails that's gonna be built in place so that information is not gonna be used to, like, train the models in any way or, like, build up the way the questions will be answered. It's just going to be spitting, like, factual information out from those those documents that have been uploaded by you.

But you always are going to be in control of what those answers will look like. You can always manually override it, as seen in the demo.

So that just means that you guys will have full control over what that answers may look like. What those answers may look like.

Mark Nafash

12:21 - 12:43

Yeah. And I think that just brings up a good point that I just wanna point out and and I think I mentioned this earlier.

Right? This is our proprietary built, LLM AI that was built with inside of SS and Cintralinks. So it does not leave our firewall.

It stays within our system. So you get the full bank rate security that comes with Intronics.

Maren Fajardo

12:43 - 13:17

It looks like there's another. How do you stop the same question being answered three different ways across three different LP requests? Yeah.

So it's all gonna be based off of those internal documents, so it likely should be the same. But, again, you will have the ability to manually override, the ability to, specify, like, firm specific terminology based off of each individual question.

So you are in control over what that answer is going to look like, but it will likely look the same, depending on what those repetitive questions do look like.

Mark Nafash

13:17 - 13:23

Looks like there's some more. Is there content used to train models or shared with other firms?

Maren Fajardo

13:23 - 13:23

No.

Mark Nafash

13:23 - 13:24

Simple? No.

Maren Fajardo

13:24 - 13:29

Yeah. Definitely not.

Mark Nafash

13:29 - 13:35

With 34 documents, how can answers be factual? Isn't the context window saturated?

Maren Fajardo

13:35 - 13:55

Yeah. So it's going to be based off of the documents that you provide.

So it's not gonna be pulling out or, like, basically having a stream of consciousness to answer those questions. It's just gonna be based off of the.

I guess that's what I meant by factuals is there's no stream of consciousness that the AI tool has when answering those questions.

Mark Nafash

13:55 - 14:39

Do you have API or MCP connections with other external source tools to enhance? Yes. We have open APIs, so you can connect with any tool that you have, that you're using today.

We do have MCP. We're in the process of rolling it out across of our product lines.

We have partners that we're already working with. So that is a great way to continue to enhance the questions.

I hear every day our clients using, different AI tools to help with DDQ. I view ours with an MCP as a great way to completely enhance and get the full picture of your responses.

Maren Fajardo

14:39 - 15:24

Does it? create, dynamic answers based on how the question is asked, or is it pulling the best answer it can find as a starting point? Yeah. So it's basically gonna be taking a first pass at answering that question.

So, depending on how that question is asked, it's going to pull specific information from from those documents. You can also choose, I'm.

I believe it was in the demo that, how. Which documents can be used to answer that question.

So if you, if you'd prefer it to be based off of two separate documents, you can do so where instead of all of the the internal AI documents that you have turned on.

Mark Nafash

15:24 - 15:47

Gesmar mentioned as well. Right? I think part of it is always having a human loop in the end to review and and make sure that you're happy with the the responses.

Any other questions?

Maren Fajardo

15:47 - 15:47

K.

Mark Nafash

15:47 - 16:25

Great. If not, I think we're gonna wrap it up here.

I really appreciate everyone's time in in listening into our webinar and answering these questions. It's super valuable to hear what you all are thinking out in the market and how we can better serve you.

If you have interest in seeing a more detailed demo, learning more about our products and fund center specifically, please do not hesitate to reach out. We're happy to spend time with any customers anywhere in the world and and further discuss these products and our road map where we're going with AI and how it can help enhance your workflows.

But, again, I really appreciate the time and the people that have joined the call today and those that are gonna listen after. Thank you.

Maren Fajardo

16:25 - 16:26

Thank you.